Improve Ceph OSD parsing reliability with caching

Replace fragile per-device ceph-volume parsing (grep -B 20) with a
single upfront query that builds lookup tables.

New build_ceph_cache function:
- Parses ceph-volume lvm list output using proper block detection
- Extracts OSD IDs by matching "====== osd.X =======" headers
- Maps block devices to their corresponding OSDs
- Queries ceph osd tree once for all status info
- Creates CEPH_DEVICE_TO_OSD, CEPH_OSD_STATUS, CEPH_OSD_IN arrays

This is both more reliable and more efficient.

Fixes: #9

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>

driveAtlas.sh

@@ -336,6 +336,67 @@ build_drive_map() {
     fi
 }
 
+#------------------------------------------------------------------------------
+# build_ceph_cache
+#
+# Queries Ceph once and builds lookup tables for OSD information.
+# This is much more efficient than querying ceph-volume per device.
+#
+# Sets global associative arrays:
+#   CEPH_DEVICE_TO_OSD - Maps device names to OSD IDs (e.g., sda -> osd.5)
+#   CEPH_OSD_STATUS - Maps OSD numbers to up/down status
+#   CEPH_OSD_IN - Maps OSD numbers to in/out status
+#------------------------------------------------------------------------------
+build_ceph_cache() {
+    declare -g -A CEPH_DEVICE_TO_OSD=()
+    declare -g -A CEPH_OSD_STATUS=()
+    declare -g -A CEPH_OSD_IN=()
+
+    # Skip if ceph-volume is not available
+    if ! command -v ceph-volume &>/dev/null; then
+        return
+    fi
+
+    # Parse ceph-volume lvm list output
+    # Format: blocks starting with "====== osd.X =======" followed by device info
+    local current_osd=""
+    while IFS= read -r line; do
+        # Match OSD header: "====== osd.5 ======="
+        if [[ "$line" =~ ======[[:space:]]+osd\.([0-9]+)[[:space:]]+======= ]]; then
+            current_osd="osd.${BASH_REMATCH[1]}"
+        # Match block device line: "      block device              /dev/sda"
+        elif [[ -n "$current_osd" && "$line" =~ block[[:space:]]device[[:space:]]+/dev/([^[:space:]]+) ]]; then
+            local dev_name="${BASH_REMATCH[1]}"
+            CEPH_DEVICE_TO_OSD[$dev_name]="$current_osd"
+        fi
+    done < <(ceph-volume lvm list 2>/dev/null)
+
+    # Skip if ceph command is not available
+    if ! command -v ceph &>/dev/null; then
+        return
+    fi
+
+    # Parse ceph osd tree for status
+    # Format: ID  CLASS  WEIGHT   TYPE NAME        STATUS  REWEIGHT
+    while IFS= read -r line; do
+        # Match OSD lines: "  5   hdd  3.63660      osd.5       up  1.00000"
+        if [[ "$line" =~ ^[[:space:]]*([0-9]+)[[:space:]]+.*osd\.([0-9]+)[[:space:]]+(up|down)[[:space:]]+([0-9.]+) ]]; then
+            local osd_num="${BASH_REMATCH[1]}"
+            local status="${BASH_REMATCH[3]}"
+            local reweight="${BASH_REMATCH[4]}"
+
+            CEPH_OSD_STATUS[$osd_num]="$status"
+
+            # Determine in/out based on reweight
+            if awk "BEGIN {exit !($reweight > 0)}"; then
+                CEPH_OSD_IN[$osd_num]="in"
+            else
+                CEPH_OSD_IN[$osd_num]="out"
+            fi
+        fi
+    done < <(ceph osd tree 2>/dev/null)
+}
+
 #------------------------------------------------------------------------------
 # get_drive_smart_info
 #
@@ -447,7 +508,10 @@ esac
 # Drive Details Section
 #------------------------------------------------------------------------------
 
-echo -e "\n=== Drive Details with SMART Status (by Bay Position) ==="
+# Build Ceph OSD cache (single query instead of per-device)
+build_ceph_cache
+
+printf "\n=== Drive Details with SMART Status (by Bay Position) ===\n"
 printf "%-5s %-15s %-10s %-8s %-8s %-8s %-30s %-20s %-12s %-10s %-10s\n" "BAY" "DEVICE" "SIZE" "TYPE" "TEMP" "HEALTH" "MODEL" "SERIAL" "CEPH OSD" "STATUS" "USAGE"
 echo "----------------------------------------------------------------------------------------------------------------------------------------------------"
 
@@ -470,31 +534,15 @@ for bay in $all_bays; do
         smart_info=$(get_drive_smart_info "$device")
         IFS='|' read -r type temp health model serial <<< "$smart_info"
 
-        # Check for Ceph OSD
-        osd_id=$(ceph-volume lvm list 2>/dev/null | grep -B 20 "/dev/$device" | grep "osd id" | awk '{print "osd."$3}' | head -1)
-
-        # Get Ceph status if OSD exists
+        # Check for Ceph OSD using cached data
+        osd_id="${CEPH_DEVICE_TO_OSD[$device]:-}"
         ceph_status="-"
+
         if [[ -n "$osd_id" ]]; then
-            # Get in/out and up/down status from ceph osd tree
-            osd_num=$(echo "$osd_id" | sed 's/osd\.//')
-            # Parse ceph osd tree output - column 5 is STATUS (up/down), column 6 is REWEIGHT (1.0 = in, 0 = out)
-            tree_line=$(ceph osd tree 2>/dev/null | grep -E "^\s*${osd_num}\s+" | grep "osd.${osd_num}")
-            up_status=$(echo "$tree_line" | awk '{print $5}')
-            reweight=$(echo "$tree_line" | awk '{print $6}')
-
-            # Default to unknown if we can't parse
-            [[ -z "$up_status" ]] && up_status="unknown"
-            [[ -z "$reweight" ]] && reweight="0"
-
-            # Determine in/out based on reweight (1.0 = in, 0 = out)
-            # Use awk for floating point comparison (more portable than bc)
-            if awk "BEGIN {exit !($reweight > 0)}"; then
-                in_status="in"
-            else
-                in_status="out"
-            fi
-
+            # Get status from cached OSD tree data
+            osd_num="${osd_id#osd.}"
+            up_status="${CEPH_OSD_STATUS[$osd_num]:-unknown}"
+            in_status="${CEPH_OSD_IN[$osd_num]:-out}"
             ceph_status="${up_status}/${in_status}"
         else
             osd_id="-"